granger vausality test in multivarate time series

hi,

i have a multivarie time series and i want to find granger causality,all variables are I(0), i have 3 questions ,

1)there is 3 ways to do that ,

• exclude-all
• leave-one-out
• block-wise

i did last one and then for specifiying the causality between variables ,i conduct second method.based on "block wise " results there is not any causality but,after running "leave on out",there was on causality .

how could i interpret this contradiction?

2) how i could run just bivariate causality ,because all of method are defined for 3 and more, is there any function except gctest that i could use?

3)may i test I(0) and I(1) simultaneously in my test (after i difference I(1) and convert it to I(0))?

Thanks in advance for your help